Re: BBC **LIES** About HIV TESTS

"Wolfgang G. Gasser" <[EMAIL PROTECTED]> schreef in bericht
news:[EMAIL PROTECTED]

>
> "Alex" in news:[EMAIL PROTECTED] :
>
> > http://news.bbc.co.uk/1/hi/talking_point/3248715.stm
>
> > From the Talking Point radio programme:
> >
> > " Robin Lustig:
> > Peter Piot there are people, aren't there, who worry that, for example,
> > tuberculosis, which can often be a manifestation of HIV infection but
> > isn't necessary so and that can muddle up the figures sometimes.
> >
> > Peter Piot:
> > Well it's true that there are - it's possible to have a so-called false positive
> > test but it's extremely rare. Also in these surveys that I just talked about in
> > Africa the same scientific standards are being applied, initial positive tests
> > are being verified with other tests. I really think that that's not the problem. "
> >
> > -----------------------
> >
> > Now let ME become a good journalist, and supply the refutation of
> > Piot's OUT AND OUT FALLACIES. From three, independent and unassailable
> > sources.
> >
> > LIE one.
> >
> > "SO-CALLED FALSE POSITIVE TESTS, BUT IT'S EXTREMELY RARE."
> >
> > Well obviously Peter Piot, director of WHO/UNAIDS is not even familiar
> > with his own organisation's data. Because in a population with a low
> > occurrence of HIV infection, false positives are THE NORM.
> >
> > 1) The WHO ITSELF.
> >
> > http://www.who.int/bct/Main_areas_of_work/BTS/HIV_Diagnostics/
> > Evaluation_reports/Operational%20Characteristics_HIV%20Report9_10.pdf
> >
> > " When a single screening assay is used for testing in a population
> > with a very low prevalence of HN infection, the probability that
> > a person is infected when a positive test result is obtained (i.e.,
> > the positive predictive value) is very low, since the majority of
> > people with positive results are not infected. "
> >
> > I will repeat for the slow of understanding that in low prevalence
> > populations using one elisa, "THE MAJORITY OF PEOPLE
> > WITH POSITIVE RESULTS ARE NOT INFECTED".
>
> [snip]
>
> > 3) HIV Screening in a Military Blood Transfusion Centers
> > This is Italian data from the general population (blood donors).
> > http://www.certi.org/CMA/newsletter/v03n01.pdf
> >
> > Table II
> >
> > Number of dontations: 25,562
> > Number of blood donations ELISA positive: 31
> > Number of blood donations after confirmation test: 2
> >
> > This means that of 25,562 people tested with a single ELISA,
> > 31 tested positive, and after further testing, only 2 *remained*
> > positive.
> > This means that in the field, ELISA had a ppv (positive predictive
> > value) of 2/31 or 6.45%.
>
> This means that there are  (31 - 2) / 25,562 = 0.11%  false
> positives.
>
> Only  2 / 25,562 = 0.008%  turned out to be HIV+.

Both of those are interesting, but they don't matter.

The question has always been how many of the people
who test positive on their first ELISA remain positive after
further testing.

The reason this is relevant, is because the "30%" number
you go on to quote, is based on these single ELISAs.

So it is irrelevant that the majority of the population in Italy
doesn't test positive on ELISA. What is important is that
93.55% who do, DO NOT REMAIN POSITIVE AFTER
FURTHER TESTING.

So, when one then goes on to South Africa, tests only pregnant
women, and then does not follow up these single positive ELISAs,
it is highly dubious that the numbers you get accurately reflect
the national HIV prevalence in that country.

As has been shown by the HSRC, who also used single unconfirmed
ELISAs (as recommended by the WHO paper I gave the URL for),
and by merely testing a representative sample of the population,
came up with not 19% HIV infection, but 11.4% HIV infection.

If these single ELISAs had been followed up with say Western
Blot testing, who is to say that number would not have been cut
in half again? Cut by half, or even more?

I find it odd that no one, regardless of the WHO recommendation
or guideline, goes on to check these single positive tests.

>
> It IS relevant whether <0.008%> or <0.008% + 0.11%>, but it is
> totally irrelevant whether <around 20%> or <around 20% + 0.11%>.

No it isn't. What is relevant is how many of the original single
ELISAs remain positive after confirmation with Western Blot.

It doesn't matter how many Italians don't test positive on ELISA,
but how many South Africans (or Italians) remain positive on
further and more sophisticated tests AFTER they test positive
on their first ELISA. And the answer at least in Italy, is _very few_.

Alex